Accessible Maps part 2 of infinity

Interactive Maps: Are they simply complex images?

In a previous post I asked “How can we make interactive maps that are more accessible?“.

We still struggle with this at work, and arguably on the internet at large. Even the biggest players, in my view, simply punt on this. ESRI tools are allegedly comformant per WCAG 2.1, US Section 508, and EN 301 549 standards. Or as they put it “Accessibility is an ongoing effort”.

But I can take a random example, in this case a map from the ArcGIS Experience Builder Project Gallery, and find a host of errors, even without considering the map:

• The big button at top right has contrast errors that fail at WCAG AA
• There is no heading structure (the main page title is a <p>)
• There is audio provided without transcription

Some of this is likely due to user configuration of the underlying tool. And sure, anyone could probably reivew my website and find issues. But also, I’m not a billion dollar company. And I’m not claiming accessibility compliance.

Anyhow, enough ranting about what’s wrong with other people’s maps. How do we do better with our own?

Maps = Complex Images, or do they?

As noted in the previous post, if you search WCAG and try to find the thing that most closely fits an interactive map, most people would argue this has to be “complex image”

But this is inadequate on many levels

• A map is not strictly an image, especially an interactive map
• A interactive map is also not a chart, although they share some characteristics
• An interactive map is really a category unto itself, it is neither an image nor a chart

Let’s say we stop here and go with the idea that we can treat an interactive map as a complex image. The general advice in that case is to provide a short and a long description, and we’re done.

From the WAI complex image tutorial (emphasis mine):

The first part is the short description to identify the image and, where appropriate, indicate the location of the long description. The second part is the long description – a textual representation of the essential information conveyed by the image.

There are no maps in that example (or any examples I can find), but let’s continue on anyway. Imagine a simple interactive map, say a choropleth showing the counties of the United States, with 5 contrast and colorblind safe colors. The map can be panned and zoomed from say z4 (1:36M) to z12 (1:144K). There is of course a basemap (there always is), and there are labels that adapt (i.e. label different things) at different zoom levels.

How are we to craft a dynamic long description that provides a comparable text-based alternative to the visual experience a sighted person derives from the map?

We do have a few options. We can ignore the the basemap and labels, and argue incidental to most relevant information that the map is meant to convey (not “essential information”, per WCAG).

And then we could, for example, desribe what we see in the viewport at each zoom level. State X has color 1, State Y has color 2, and so on. This could be automated without too much difficulty. This could even be provided as a long description.

And it could be provided in tabular form, in compliant HTML tables, or as direct data downloads.

But have we achieved accessibility? Is a desription or tabular presentation of the data in this way truly comparable to what sighted users get from the same map? I think immediately of the visual patterns that even the most basic map provides. A concentration of values in certain areas or regions of the map, and a paucity of values in other areas. Adjacency of similar gradients colors in some areas, sharp contrast in others areas. This is also part of what the map “says”, and how can we achieve that with text, in an automated way, across all possible views?

What is the way forward?

I honestly do not know. What I do know is that the existing treatments and solutions, e.g. short and long-descriptions seem wholly inadquate.

It would be great if WCAG had additional resources that address this directly for interactive web maps, perhaps by splitting maps out of the “complex image” trap entirely, into their own separate category, with advice and examples to match. What shape this takes, how we actually achieve map accessibility, remains to be determined.